Speed test Servers

Blessings for 2014 to one and all. Can anyone recommend a good speed test server that would facilitate a Fiber Speed Test?

Thanks in advance.

Thinkbroadband?

Use this http://speedtest.btwholesale.com/ as I find it spot on all the time with ADSL, Cable and VDSL reliable speed test.

Edited by adslmax (Thu 02-Jan-14 13:55:55)

Doesn't that depend on ISP as I believe it is only available for BT wholesale connections? OP appears to be on Cable and Wireless (isn't that Virgin now? EDIT - appears not but not sure what bearers they use).

EDIT again : Seems I was wrong as that test seems to work from work connection which is commercial leased line (although is getting nowhere near the speeds the line is capable of).

Edited by ian72 (Thu 02-Jan-14 14:01:06)

Thank you for the kind words.

http://www.speedtest.net

The London servers Namesco, Netcetera and actually Virgin Media seem to work quite well.

http://www.speedtest.bbmax.co.uk/

Is another Namesco speed test.

Beyond that, if your ISP has their own speed test server try and use that where possible to eliminate the Internet from the equation.

My virgin media speedtest was spot on using bt speed test
http://s5.postimg.org/u15spd0kn/speedtest_on_cable.jpg

This speed test is also spot on too: http://fibretest.kc.kcom.com/ (u need a java to enabled for firewall, ping test)

In reply to a post by DonnyS:

Blessings for 2014 to one and all. Can anyone recommend a good speed test server that would facilitate a Fiber Speed Test?

Thanks in advance.

http://www.thinkbroadband.com/speedtest.html

http://www.thinkbroadband.com/speedtest/button/13886...

Apart from the speedtest.net calculation methodology which is sometimes a little suspect you need to read up about it.

I tend to get consistency from individual speedtest.net sites but they rarely match each other even when located in the same city/region - upstream can be 5, 10, 12, 15 or 18 Mbps depending on server. So, if you use one, stick with it and for comparison only.

in the case of your slow upload results its likely BDP related. So limited by latency/sendbuffer.

Unlikelly ... two speedtest.net servers located in the same area and with similar latencies. One gives 56 down and 5 up whilst the other 62 down and 18/19 up. Neither of which is totally correct - based on sync speed I should get 56 down 19 up. Nothing changes here just the routing from Telecity/Telehouse

I agree with MHC that different speedtest.net servers give very different results.

I have found the thinkbroadband test most reliable for upload speeds; it is the only one recently that gives results that match the real speeds of good backup sites (eg SquirrelSave) at around 18Mbps (visualware test used to but went seriously bad some months ago).

For me, the thinkbroadband TBBx1 download test always gives really low results (9.3), and the HTTPx6 test usually gives slightly low results (but quite good ones this time, 66.3)

http://www.thinkbroadband.com/speedtest/results.html...

Edited by StephenTodd (Thu 02-Jan-14 16:58:16)

give me the 2 servers I will test as I have a large send buffer configured.

Also you may have failed to take into account the receive buffer on the speedtest server.

Edited by Chrysalis (Thu 02-Jan-14 17:43:39)

In reply to a post by StephenTodd:

I agree with MHC that different speedtest.net servers give very different results.

I have found the thinkbroadband test most reliable for upload speeds; it is the only one recently that gives results that match the real speeds of good backup sites (eg SquirrelSave) at around 18Mbps (visualware test used to but went seriously bad some months ago).

For me, the thinkbroadband TBBx1 download test always gives really low results (9.3), and the HTTPx6 test usually gives slightly low results (but quite good ones this time, 66.3)

http://www.thinkbroadband.com/speedtest/results.html...

same test with good buffer sizes?

http://www.thinkbroadband.com/speedtest/results.html...

Try Maidstone and Canterbury - 8 Mbps and 19 Mbps

And I have no control over te receive buffer and if a site is going to host a service then it should be correctly configured.

In reply to a post by Ignitionnet:

http://www.speedtest.bbmax.co.uk/

Is another Namesco speed test.

I just love the "Get Faster Broadband Here - Download speeds up to 8Mbps available!" that still appears on that one.

maidstone has a poor recv buffer maxing out at about 17.5mbps for me and canterbury 19mbps.

Plusnet congestion tho slowing down downstream on both.

Proves the point that not all speedtest.net servers are fully capable of working with 80/20 services or full rate FTTP.

In reply to a post by MHC:

Proves the point that not all speedtest.net servers are fully capable of working with 80/20 services or full rate FTTP.

A lot of them I think are overloaded. Matt from Uno certainly said theirs often had a LOT of load a year or so ago.

In reply to a post by Chrysalis:

Plusnet congestion tho slowing down downstream on both.

Generally how have you found PlusNet compared to BT Infinity? I'm starting to look at alternatives, as I want static IP again, but I don't like the PN traffic management.

(I have big issues with person A's legit traffic being slowed down because person B wants to watch iPlayer, and think all people paying the same should see same slowdown).

In reply to a post by jchamier:

(I have big issues with person A's legit traffic being slowed down because person B wants to watch iPlayer, and think all people paying the same should see same slowdown).

But that is the real issue, if you slow down persons B's traffic they really will see the slow down.

As you are probably aware I'm with PN, and I have used torrents on a Sunday evening and noticed no slow down, in fact the torrent would almost max out my connection, although I'm only getting around 47Mbps.

I would also be interested to hear what Chrysalis thinks so far.

Edited by R0NSKI (Thu 02-Jan-14 20:16:23)

In reply to a post by R0NSKI:

But that is the real issue, if you slow down persons B's traffic they really will see the slow down.

I just don't think its fair in pricing terms. That my streaming a pay per view movie to my Apple TV is slowed down in preference to someone watching iPlayer. (Since the ISP can't tell if I'm streaming the Apple movie, or downloading it to watch later on iPad/iPhone or Mac/Windows PC).

As you are probably aware I'm with PN, and I have used torrents on a Sunday evening and noticed no slow down, in fact the torrent would almost max out my connection, although I'm only getting around 47Mbps.

I hope being in BT ownership that the actual effects of the traffic management are mostly non visible to everyone.

I would also be interested to hear what Chrysalis thinks so far.

maybe? is your line interleaved.

dispite all the comments on forums about interleaved having no negative affects on speeds, it actually does when you understand BDP.

A big factor tho more so than servers is the derfault configuration on client operating systems, out of the box windows 7 eg. cannot handle much more than 3-4mbit/sec per thread on 30ms latency.

So far it isnt all smooth sailing. It also isnt terrible tho.

I think 2 days is a bit early for a reasonable review so consider this more "notes"

PPP dropped by plusnet first morning after connected, which has me wary of frequent disconnections, I wasnt disconnected this morning.
Some visible congestion this evening, the TBB speedtester shows 50-75% throughput on the non threaded test, on BT that was maxed out everytime.
On external ftp and some other tests I did from fast sources some light visible congestion.
Some light congestion visible on speedtest.net.
Before I upgraded to pro I had packetloss on ssh sessions using non standard ports.

Its a list of things, but I will point out the congestion is light, not heavy.

No,

And I fully understand interleaving - I was doing teh equivalent of mux'ed digital voice comms back in the late 80's.